Transcript for:
Stenography and AI Speech Recognition

Ladies and gentlemen This meeting...... Why do people hire us stenographers? Because we can be faster and more accurate A three-hour meeting We can publish the manuscript on the spot Machines must be faster than people A three-hour meeting We only need 7 minutes to complete the transfer I am Honghong (Nanjing, Honghong, Senior Stenographer) My name is Wang Qingran (Hefei, Wang Qingran, AI Speech Recognition Programmer) 27 years old this year My position is artificial intelligence pogrammer (Wang Qingran, 27 Years Old, Hefei, AI Speech Recognition Programmer) I am now in Hefei (Wang Qingran, 27 Years Old, Hefei, AI Speech Recognition Programmer) I'm a senior stenographer (Honghong, 27 Years Old, Nanjing, Senior Stenographer) Now in Nanjing (Honghong, 27 Years Old, Nanjing, Senior Stenographer) My job is to record sound THIS IS US I usually put on my makeup for 5 minutes I'm pretty nervous about the meeting later Because the leaders of the National Development and Reform Commission and the State Council are coming It's a advanced meeting Simply A stenographer is to convert all sounds that can be changed into words into words Market research Round table Seminar Stenographers are standard Into the industry It's also quite a coincidence There is such a course in college After graduation, I went to Beijing shorthand association Studied for a while During this period I interned at the State Department and the Ministry of Defense The meeting starts at 9 o'clock in the morning I usually arrive before 8:30 There are still 4 stops in the past In fact, the meeting in Nanjing The traffic is quite convenient Basically it will be there in about half an hour Nanjing is a central city It can cover some stenography needs in northern Anhui and northern Jiangsu For me, it is also a convenience in life The best of both worlds Help me navigate to iFlytek iFlytek Voice Industry Base Located near Yonghe Road, Shushan District 9.5 kilometers Our company's work time start at 9:30 I usually leave at this time Then half an hour to the office Yunmi refrigerator You can punch in when you approach the company in the morning I am a computer major from Tianjin University When I was in school, I came into contact with some directions of artificial intelligence speech recognition technology With some foundation After graduation, I came to iFlytek My family is Hefei And iFlytek is in Hefei Worked for more than three years Has been conducting research related to speech recognition Recent projects are mainly Optimize speech recognition rate in conference situations Our speech recognition field One of the most successful situations now applied Is the meeting scene The meeting I attended this time is about High-level seminar on environmental protection and low carbon Generally after entering the venue I connected my audio cable to the console first The purpose is to make the sound clearer After sitting down, write down the words that may appear at high frequencies For example, high-quality development Fourteenth Five-Year Plan Double carbon I can list This way you can operate more quickly Still a little nervous It is after all too professional meeting Environmental requirements I think people may be more adaptable than artificial intelligence Then more flexible We can actually use multiple channels To collect some noise counterexamples To improve the accuracy of speech recognition For example, in our meeting scene It might make some chair-pulling sounds Door closing sound Coughing sounds Hello x4 Open the window The audio we collected in this car is actually Some noise data And some command data Navigate to iFlytek It will capture the speed and tone of my speech And the actual distance We will add it to the training set when we go back to train it Distinguished guests This meeting is about to begin Please sit down as soon as possible And turn off cell phones and other communication devices Or put it on silent Thank you for your cooperation Ok yes Now let's talk briefly Yangtze River Delta Chemical Industry Zone At the same time, there are companies of a certain size in our industry In the past 5 years, there have been 3,700 fewer...... Our instantaneous speed will reach 500 or 600 words (per minute) But your average is basically two or three hundred words (Per minute) They can type 600 words in a minute?! How can a stenographer type out a text after finishing every sentence The speed recorder is specially designed for typing The function is pure The speed recording machine has a total of 24 keys Recompiled based on Pinyin There are 12 key positions on the left and right Axisymmetric And exactly the same Take the word meeting as an example Normal computer keyboard We need to hit 5 keys back and forth to type Stenographers can drop keys with both hands at the same moment You can type it with a tap Further accelerate the ecological green integration demonstration zone Shanghai Pilot Free Trade Zone Lingang New Area This feeling is a very typical Conference transcription scene (Ran is processing the audio of the meeting) Now the audio it has begun to process Stat process Decoding is a bit slow So while waiting for the program to run You can also drink a cup of milk tea This is quite pleasant For a human stenographer they must concentrate on 2-3 hours of meeting time to listen to write Harder Speech recognition You trained this model well Enter the audio It will give a stable result Faster and once and for all Actually, I'm done here This is the manuscript we transferred It will have some keyword errors 2030 Carbon Peak (See the wind) 2060 Carbon Neutral (See through and) Both places It identify the wrong words Many catchwords Names of people and places that have not been seen We will recognize wrong But I feel that human words should recognize At the request of the general secretary Development of the Yangtze River Economic Belt Our Carbon Peak, Carbon Neutral demonstration area (Carbon Peak, Carbon Neutral) What's next? Our stenographer's job It's not just about recording each words We want to make the manuscript readable and logical When a guest's sentence is not smooth I'll change the order or complete it The second aspect is I will follow the guest's logic And show the PPT to list 1234 This requires a stenographer discerning, organize, summarize We can't be like stenographers Go make some changes to this text It is a very faithful Show every word you say But speech recognition has actually developed in recent years Not the same as in previous years We will make a punctuation prediction Tone prediction Oral regularity Transition of paragraphs If he really succeeds Then there is really no value for our existence This concludes our morning meeting Thank you all Now in the final proofreading See if there are any obvious typos Then it can be sent to customers immediately Except the behind-the-scenes workers We are often the first to enter the conference room And the last one to leave the conference room On average, there are 20 meetings a month The pay for a day in Nanjing is about Around 600 yuan (1200 Yuan per month) Our industry 4~ 50,000 yuan for first-tier cities Second and third tier cities 2~ 25,000 yuan On average, there are 20 meetings a month But there may be half-day meetings in these 20 meetings Also whole-day meetings Time is free All sold out But our industry is quite lonely A stenographer is a loner Go alone and come back alone There is no difference between weekdays and weekends Sometimes I often stay at home and don't want to go out to play Because when I have time, others may not have time I don't have time when others have time Because now I want to take the junior accounting qualification certificate So I'm watching the teaching videos Take the time to improve my other skill levels Ps And the accounting I'm studying recently Make life a little fuller A little richer Not just stenography skill The intensity of our working hours is actually quite large Until around eight or nine o'clock in the evening Hanger, turn on the lights Hanger drop Hanger lights out My work as a programmer The biggest impact on my life is Make the whole person more rigorous Like playing games I won't say use a general thing to Evaluate one thing I won't say about playing games at night Until I want to sleep, then I will stop it I would definitely say I go back tonight I play games for an hour I will definitely stop the game after an hour I will do something else Stenographer in peacetime Listening to so many people at work Do you not like to hear others talk in life Yeah Just like to stay alone kind of End a meeting If I'm not in a hurry I will go for a walk in the nearby park This time is the heart of nothing else That is Zifeng Building It is the largest tall building in Nanjing Often go to the Zifeng Building for meetings When I went to an Internet conference a few years ago There is artificial intelligence show up I had self-doubt I wonder if my work is worthless At this stage I think I should let myself be at peace All kinds of meetings make me feel that my vision has become wider I can see a different kind of scenery See different people Keep fresh at all times Is what makes me like this industry very much The most challenging thing about the job How to put our technology into practice into what everyone really needs Artificial intelligence I think the original intention of its development Not to replace humans (2012 Voice Evaluation; 2015 speech recognition; 2017 Smart Medical Assistant Robot) But to help humans work better (2012 Voice Evaluation; 2015 speech recognition; 2017 Smart Medical Assistant Robot) HI Thank you for seeing here I am the Producer of this episode, Leying. This is also the first video I made when I came to DX Before making this video I thought the two professions in the film It must be the binary opposition between replacement and replacement But after I came into contact with Honghong and Ran I found a lot of overlap and integration between their careers The profession itself has its own advantages and disadvantages And their own unique living space What I admire more is their mentality Honghong relaxes her mind under the sense of crisis Upgrading skills Found her own peace And Ran is standing in the perspective of people-oriented Dialectical treat technology Harvest the sense of accomplishment of constantly breaking through himself I wonder what you think of these two professions Have you ever encountered Worried about being replaced by technology Welcome to tell us in the comment And finally, if you like today's video Don't forget to share this video Subscribe our channel Turn on the notification See you next time, bye~